[klibc] [klibc:master] arm/setjmp.S: fix longjmp

[Steve McIntyre]

On Sun, Jul 22, 2012 at 03:50:42PM +0000, Thorsten Glaser wrote:
>H. Peter Anvin dixit:
>
>> klibc explicitly considers longjmp(..., 0) to be user error
>
>This IMO is a possibly dangerous deviation from guaranteed¹
>behaviour, and hard to track down. Also, checking this in
>longjmp asm probably is easier than to offload this to all
>users.

I'd be surprised to see that, myself. The point of my patch was at
least to give consistent behaviour...

-- 
Steve McIntyre, Cambridge, UK.                                steve at einval.com
"You can't barbecue lettuce!" -- Ellie Crane